Trajectory Tracking With Visual Localization and Odometry Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indoor navigation technologies face challenges in accurately tracking trajectories in new environments due to accumulated errors from odometry and high computational loads in visual localization, making precise positioning difficult and time-consuming.
Innovation Solution
A method that combines visual localization (VL) and odometry technologies to optimize trajectory tracking by correcting relative poses using absolute poses estimated from VL, minimizing errors through a pose graph optimization algorithm, and utilizing visual object tracking for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If odometry is used for trajectory tracking, then real-time positioning is achieved, but accumulated errors occur over time
Solution Approach 1:
The system implements feedback by using visual localization to detect accumulated errors in odometry-based trajectory tracking and correcting them in real-time. The visual localization provides absolute position references that feed back into the odometry system to eliminate drift, thereby maintaining both real-time performance and long-term accuracy.
Solution Approach 2:
Visual localization serves as an intermediary system that mediates between the odometry-based continuous tracking and absolute position references. It translates visual features into position corrections that are applied to the odometry trajectory, bridging the gap between relative and absolute positioning without requiring complete system re-initialization.
2Measurement precision
If visual localization is used for trajectory tracking, then positioning accuracy is improved, but computational load increases
Solution Approach 1:
The system applies partial visual localization by using it only at specific intervals or keyframes rather than continuously processing visual data at full resolution. This selective application provides sufficient positioning accuracy corrections while significantly reducing the computational burden compared to continuous full-scale visual localization.
Solution Approach 2:
The trajectory tracking is segmented into odometry-based continuous tracking segments and visual localization correction segments. By dividing the positioning task into these alternating phases, the system maintains real-time performance during odometry segments and applies computational-intensive visual localization only when needed for error correction, reducing overall computational load.
3Measurement precision
If standalone visual localization is used, then absolute positioning is achieved, but it is time-consuming and requires high computational resources
Solution Approach 1:
The system performs preliminary action by pre-processing and storing visual features and descriptors in advance during map creation or initial exploration phases. When trajectory tracking is needed, the system queries these pre-processed features rather than performing complete visual localization from scratch, dramatically reducing positioning time while maintaining absolute positioning accuracy.
Solution Approach 2:
The system merges visual localization with odometry by combining their respective strengths. The odometry provides continuous relative motion tracking while visual localization provides periodic absolute position corrections. This merging allows the system to achieve accurate absolute positioning without relying solely on computationally intensive standalone visual localization at every time step.
Data Source
AI summary
The trajectory tracking method for a mobile electronic device may include tracking a trajectory of the electronic device by using results of pose estimation using odometry and results of pose estimation using visual localization (VL) as camera pose information.